“Holden-ball” 2020/21
Focus of this piece:
A reflection on last season, but mainly based on comparing Holden to Johnson data, as there was a marked change in system/approach (more on this later), and how this is expected to follow into this season (2020/21).
Season 2019/20 Summary:
A mixed season that ultimately ended up with a 12th placed finish, and an unremarkable 17 wins, 12 draws and 17 defeats. This from a side that spent a third of the season occupying a top 6 (playoff) position, and until the sacking of Lee Johnson after 41 games following a defeat versus Cardiff, was still in the mix. Dean Holden (assistant head-coach) took over as caretaker head-coach for the remaining 5 games, earning 8 points, but the damage had been done, especially either side of the Covid break where City picked up 2 points out of a possible 27. Holden has since been appointed as permanent head-coach, bringing in Paul Simpson and Keith Downing as assistants, both having success with England age-group sides.
Recruitment Summer 2019:
Ignoring the timing of last summer’s recruits, nobody can say Lee Johnson wasn’t backed. Jay Dasilva (LB), Tomas Kalas (CB) and Kasey Palmer (AM) all returned from Chelsea, but on permanent deals. Dan Bentley (GK), Sam Szmodics (AM), Adam Nagy (CM) Han-Noah Massengo (CM) all arrived for fees from Brentford, Colchester, Bologna and AS Monaco respectively. Tommy Rowe (LM) and Rene Gilmartin (GK/coach) were picked up on free transfers together with out of contract Rodri (CF) and Ashley Williams (CB) after the window closed. These permanent signings were supplemented by Pedro Pereira (RB) and Benik Afobe (CF) on season-long loans.
As is the Bristol City way, players were sold for top of the market fees, notably Marlon Pack, Adam Webster and Lloyd Kelly, but also Mo Eisa.
Pack’s sale, in particular, was a surprise, as Johnson had evolved a system with Pack at the heart of it, as was Jamie Paterson’s loan to Derby. Neither Massengo or Nagy were similar to Pack, and it proved evident over the season, especially when Josh Brownhill went to Burnley in January, that from a fan’s perspective recruitment had been disjointed.
Recruitment January 2020:
In January, City signed Nahki Wells from QPR and brought in Markus Henriksen and Filip Benkovic on loan from Hull City and Leicester City. However, for 3 years running City fans saw City tail off with the new signings failing to integrate ready for a playoff push. Paterson returned from Wayne Rooney’s Derby County.
Tactical System:
Probably the most frustrating aspect for me during Johnson’s reign was the lack of a clear system for a sustained period of time. There was a view amongst pundits last season that Lee Johnson’s Bristol City were still playing high press, pass through the lines football that was seen in parts of the 2017/18 season, highlighted by media coverage of the EFL Cup run. That system disappeared towards the end of that season, and in 2018/19 we started to see a system based on Pack as the quarterback. Performances started to drop off even if results were still top 10. So, the start of the 2019/20 season saw a 4231 at home to Leeds. It looked like recruitment was kind of aligned. A 3-1 defeat to Leeds, followed by the sale of Pack and injury to Jay Dasilva saw the 4231 and pre-season drilling tossed into the bin!
A 5212 with wingbacks evolved, with Kasey Palmer enjoying a free role and Andreas Weimann and Benik Afobe revelling in his clever passing. It was not to last long, as Afobe ruptured his ACL after 5 games (3 wins and 2 draws), and without bringing an abrupt end to this section, Johnson failed to find a system and players to fit for the rest of the season, flip-flopping between a back 3 and back 4, 1 or 2 up top, often all in the same 90 minutes!
The first thing Dean Holden did as caretaker was to go back to a back 3 with wingbacks. The midfield 3 was fluid with Korey Smith, Andreas Weimann and Jamie Paterson, and the recall of Filip Benkovic to the team allowed the team to build from the back and through the midfield. Fans had been crying out for a 3-man midfield, but most were shocked that it comprised a Number 10 (Paterson) and a Forward (Weimann)!
High-Level Observations:
Nothing rocket science in the following statement – City didn’t create enough chances and allowed the opposition too many. Many statistical sources suggested City’s results were better than their xG (expected goals) and league position suggested. This eventually caught up with City either side of the COVID break.
The following charts (pic 1 and pic 2) show results based on shots versus xG, both for (attacking) and against (defending). It is fairly easy to see that in the games City had more shots and / or xG, they picked up better results. In the games where City conceded more shots, xG was also generally higher and results suffered. There were only 5 games where City’s shots and xG bettered the average Shots against and xG against (see pic 1 yellow shaded area).
In pic 2 you will see there were many games where the opposition took more shots and had higher xG than City’s attacking average.
A combination of these meant eventually results came home to roost.
Pic 1: Attacking – Shots For versus xG - (green = win, grey = draw, red = loss, square size = number of goals)
Pic 2: Defending – Shots Against v xG Against - (green = win, grey = draw, red = loss, square size = number of goals)
Holden v Johnson
Although the number of games each had in charge (Johnson 41, Holden 5) are disparate, there are some clear trends even with the very small sample size for Holden. Those 5 games included a mix of opposition styles and quality to enable a reasonable view:
§ Hull City – direct
§ Middlesbrough – direct
§ Stoke City – mixed
§ Preston – mixed, high tempo
§ Swansea – passing
There is a view from some fans that the 5 games were “dead rubbers”, but Hull, Boro and Stoke were fighting for Championship survival, whereas Preston and Swansea were fully in the play-off mix, as were City after Holden started with 2 wins. Throw in the Warnock factor at Boro and the Preston jinx and I think Holden did well to come through with 2 wins, 2 draws and 1 defeat in 5 games.
Let’s not forget that City were on a 9-game winless run at the point Johnson was sacked, the worst form in the Championship at the time. Even Hull, in freefall at the same time, were in better form!
Although easy to look at averages and draw conclusions, much of the detail and insight is in looking at the spread of data attributes game by game.
As mentioned earlier, Holden stuck to a 532 for each game and pretty much the same starting eleven except where injuries necessitated change.
I’ve broken the data down into 3 main areas:
§ Passing / Possession
§ Goals / Attacking
§ Winning the Ball
Passing / Possession:
Under Johnson, City failed to control possession and were more than happy to play on the counter-attack. Holden instantly changed to a more possession-based system, and for the first time in a good while, with obvious patterns of play.
Pic 3: Passing / Possession
The first thing to pick out is the increase in passing numbers in the “All Passes” row, Total passes, 440 vs 351, (up 25%). This continues with successful passes, 361 vs 273 (up 32%) and percentage successful, 82% vs 77%. You can see the spread of games under Holden, and even lower passing games versus Hull and Boro (both very direct styles), show that passing was still better than many games under Johnson.
We can also see that Possession has increased, 56.3% vs 47.5% (up 18%). This is a significant shift, and shows Holden is very keen to ensure we take care of the ball. City had more passes before the opposition either tackled, intercepted, challenged or fouled. They moved the ball that bit quicker and crisper utilising Benkovic in the defensive line and Smith and Paterson in midfield. Without the ball there was little change, and I will cover this more in “Winning the ball” later.
Okay, but any team can increase their passing stats knocking it around at the back, and not advance the ball up the pitch, can’t they? To some extent we saw that trait under Johnson, with the defence knocking it around until they were pressed and aiming a direct ball to Diedhiou. But that was not the case with Holden (see pic 4).
Pic 4: Passes with attacking intent
We can see in each metric – Forward Passes, Progressive Passes and Smart Passes that under Holden we made more purposeful passes and made them with greater accuracy. Players like Diedhiou who has at times been guilty of losing possession with poor control, was now timing his movement to come short and either lay the ball off, or spin / roll his marker and make ground before moving the ball on.
Goals / Attacking (and a bit of Defending too):
This is probably an area where we didn’t see much disparity in the numbers between Holden and Johnson’s reigns. For me, the difference was that under Holden we got into better positions in the final third more regularly but were ultimately wasteful with the final ball or shot.
Pic 5: Goals (per game) for / against
A slight improvement in goals for and against, but no clean sheet. A missed penalty by Diedhiou versus Swansea meaning Holden’s first defeat, showed how he can use substitutions to influence a game, something Johnson did quite successfully in the past.
Pic 6: Attacking
We start to see a shift away from Johnson’s counter-attacking under Holden, as City achieved more positional attacks and got a better percentage of shots away. City still counter-attacked where appropriate. xG (expected Goals) show a small increase, as does shots, but no better accuracy. A sure sign that improvement is required in the final third this season coming. If you marry this up with much better numbers in the infamous “box entries” (Penalty Area Entries) and touches in the box, it further points to needing to be much better in this key area of the pitch going forward. Other teams are ruthless.
Nahki Wells form was very mixed since his arrival late in the January window. Confidence appeared low and Johnson had been critical of him on more than one occasion to the press. In the last 5 games, he appeared more confident, getting into better scoring positions, even if his link-up play wasn’t quite up to scratch. Two goals at Boro showed his quality, and his ability to stretch defences in behind gave City’s midfield that bit more space and time.
In terms of the opposition, if we go back to my high-level observations (see page 2), Holden needs to find the right balance to reduce the number of shots on the City goal. Shots against in 2 games (Preston and Swansea – see pic 8) are in the high teens, no coincidence these were the two teams challenging for the playoffs. Good teams have found ways to create chances against City under either head-coach. As a fan, you always think the opposition can hurt us, no game is safe.
Pic 7: xG against / Shots against
Pic 8: Shots against (Preston – left, Swansea – right)
Winning the Ball
On the face of pure numbers, the Johnson side appears more combative (202 vs 222, 10% down), but if you consider the possession shift from 47.5 / 52.5 (Johnson) to 56.3 / 43.7 (Holden), it’s hardly surprising that the Holden team needed to defend less often. We see a small upside in percentage of Aerial duels won, possibly down to the inclusion of Benkovic, but possibly down to the aerial bombardment from Hull and Boro. Under Johnson, City had to withstand 93 aerial duels against Cardiff (a). The constant in the Boro and Cardiff games – Warnock!
We do see some better variance when we start to look at the Offensive and Defensive duel breakdown. Johnson’s team more active in each box, but we can draw from this that Holden’s team controlling the middle portion of the pitch.
Pic 9: Duels
In terms of winning the ball back its pretty even between Holden and Johnson (pic 10), but we start to see improvement in Ball Losses, where under Holden City lose a lot less in their own defensive third, and less in the middle third. This ties back into possession and passing improvements, showing more care for the ball. Finally, sliding tackles much lower under Holden, showing less desperation, but Interceptions too, again pointing to less necessity.
Pic 10 – Winning it / Losing it!
City performances under Johnson tended to be poor when the side got “strung out” (large distance between defence and attack) leaving us exposed. From a simple look at player average positions in Wyscout maps we have seen the final games under Johnson having 45 yards front to back, whereas under Holden we saw a more compact 35 yards. City were able to get back into shape better defensively and rarely looked exposed on the break.
Players Used:
Pic 11: Appearances under Holden (red = start, yellow = sub, grey = unused sub)
§ Bentley, Dasilva, Paterson, Weimann and Wells started all 5 games under Holden
§ Hunt started 4 until getting a knock v Swansea (a)
§ Smith started 4 until his contract situation meant he asked not to play v Preston
§ Vyner appeared in all 5 matches, adding flexibility at RB, RCB and RWB and doing his career at City no harm at all
In summary, Holden preferred to pick a settled side, only really making changes where necessary and didn’t deviate away from a 352. Benkovic was recalled having featured sporadically under Johnson.
Season 2020/21 – under Dean Holden
Recruitment:
Firstly, let’s look at the players who’ve left:
§ Matty Taylor – out of contract (Oxford United – where he’d been on loan)
§ Bailey Wright – out of contract (Sunderland – where he’d been on loan)
§ Korey Smith – out of contract (Swansea City)
§ Ashley Williams – out of contract
§ Niki Maenpaa – out of contract (although rehab at City as injured whilst here)
§ Rory Holden – sold (Walsall)
§ Pedro Pereira – loan end (Benfica)
§ Benik Afobe – loan end (Stoke City)
§ Filip Benkovic – loan end (Leicester City)
§ Markus Henriksen – loan end (Hull City – now released)
No unexpected departures as it stands, unlike last summer. Covid influenced perhaps? Several u23s will be loaned out. So far Jonny Smith (Swindon Town), George Nurse (Walsall), Robbie Cundy (Cambridge United) and Saikou Janneh (Newport County) have had season long loans agreed (all have January recall options).
As it currently stands, Holden has signed one player - Joe Williams (Wigan Athletic) for £1.25m. A box-to-box, combative midfielder, who can play - this type of midfielder has been on the list of most fans for a long while. Holden has filled that gap instantly. Williams also featured for England u20s under assistant Downing.
But where the midfield is further strengthened is by the returning loanees - Joe Morrell (Lincoln City) and Liam Walsh (Coventry City). Walsh had a stellar season for promoted Coventry and there are high hopes that he’ll prove as creative in the Championship. Morrell, an Academy product, had a very impressive season at Lincoln, together with breaking into the Welsh national team and playing in two key qualifiers.
Other returning loanees include Taylor Moore (Blackpool), Antoine Semenyo (Sunderland) and Sam Szmodics (Peterborough). Moore has a big chance to become one of the ball-playing Centre-Backs in Holden’s set-up.
Squad Depth:
Pic 12: Depth chart (age in brackets, Championship appearances alongside)
At the time of writing it looks as if Pedro Pereira could be returning on a season’s loan too. That would give the RWB slot some competition and allow Vyner to concentrate on his development at RCB. The squad appears to have much better balance when looking at it from a 352 perspective, although an additional CB (left sided preferably) is needed. Another striker may depend on the futures of Szmodics and Semenyo. Other than these two, City are sorted.
Tactical System:
As mentioned throughout this piece, Holden is setting City up to play 352. From his first interview in role he mentioned that he thought the players were suited to this system, in fact early in the previous season Johnson was using it, so the fundamentals were / are there.
There have been the obvious clichés of attacking football, press from the front and pass through the lines, but it has been evidenced to an extent in the 5 games as caretaker. As for pre-season, the media team have kept things close to their chests but the word from the local press when allowed to attend has been more 352.
The question on many fans lips is what will be the midfield 3. With lots of options, including the use of Weimann last season, the permutations are numerous. With the signing of Joe Williams, many assume he will be the anchor, but he has a greater range than sitting in front of the 3 Centre-Backs. The beauty of the City midfield, albeit short in Championship experience (see pic 12) and inches in some cases is that they have many facets to their game. Through pre-season we’ve seen plenty of rotation, but it does look like Weimann will find himself back in the forward line. Paterson looks like he will remain key to breaking the lines, either with his dribbling or passing, but Niclas Eliasson has also featured in the midfield as Holden looks to go wing-less.
Up top, Holden will more than likely perm 2 out of Wells, Diedhiou and Weimann, with Wells being the main goal hope. Semenyo and Szmodics will need to provide cover, although Szmodics’ energy may afford himself more minutes as the one to re-enforce the press late on in games, similar to the role Matty Taylor used to provide. During pre-season we have seen Wells making early runs and there appears to be a clear instruction to pass it forward. All of his goals in a hat-trick versus Cheltenham Town involved running off the shoulder of his marker somewhere in the move and getting an early pass (Bakinson twice, Eliasson once). The same was true for his assist for Weimann as City completed a 4-0 win.
Bentley | Hunt, Vyner, Kalas, Baker, Dasilva | Bakinson, Massengo, Eliasson | Weimann, Wells
Not used: Wiles-Richards, J.Smith, Janneh
Other pre-season games:
§ Newport County 1-0 (Wells pen) – media blackout
§ Bristol City (mixed with u23s) – 2-1 (Szmodics 2) – media blackout
§ Aston Villa 1-1 (Paterson) – Two 60-minute halves
1st: Bentley | Hunt, Kalas, Moore, Baker, O’Dowda | Morrell, Eliasson, Paterson | Diedhiou, Wells
2nd: O’Leary | Nagy, Vyner, Bakinson, Moore, Rowe | Massengo, Szmodics, Palmer | Semenyo, Weimann
§ Swindon Town 1-1 (Weimann)
Bentley | Hunt, Vyner, Kalas, Rowe, O’Dowda | Morrell, Nagy, Paterson | Diedhiou, Weimann
60 mins: Bentley | Vyner, Moore, Kalas, Rowe, Dasilva | Palmer, Massengo, Eliasson | Weimann, Wells
Not used: O’Leary, Bakinson, Szmodics
So, there we have it, an intention to play from passing football from the back, through the midfield. The 5 games as caretaker weren’t perfect, but the signs were encouraging, the numbers back it up. Holden has now had a pre-season to help drill those with and without the ball systems. Reasons to be hopeful, as City fans its sometimes all we have!
Holden-ball here we come!!